 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ARTICLE |
a Divisions of Pediatrics
b Child and Adolescent Psychiatry, Department of Clinical Sciences, University Hospital, Umeå, Sweden
c Section for Pediatrics, Department of Women's and Children's Health, Uppsala University, Uppsala, Sweden
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
METHODS. We examined 11-year-old children who were born before 26 completed weeks of gestation in Sweden between 1990 and 1992. All had been evaluated at a corrected age of 36 months. At 11 years of age, 86 of 89 survivors were studied and compared with an equal number of control subjects, matched with respect to age and gender. Behavioral and emotional problems, social competencies, and adaptive functioning at school were evaluated with standardized, well-validated instruments, including parent and teacher report questionnaires and a child self-report, administered by mail.
RESULTS. Compared with control subjects, parents of extremely immature children reported significantly more problems with internalizing behaviors (anxiety/depression, withdrawn, and somatic problems) and attention, thought, and social problems. Teachers reported a similar pattern. Reports from children showed a trend toward increased depression symptoms compared with control subjects. Multivariate analysis of covariance of parent-reported behavioral problems revealed no interactions, but significant main effects emerged for group status (extremely immature versus control), family function, social risk, and presence of a chronic medical condition, with all effect sizes being medium and accounting for 8% to 12% of the variance. Multivariate analysis of covariance of teacher-reported behavioral problems showed significant effects for group status and gender but not for the covariates mentioned above. According to the teachers’ ratings, extremely immature children were less well adjusted to the school environment than were control subjects. However, a majority of extremely immature children (85%) were functioning in mainstream schools without major adjustment problems.
CONCLUSIONS. Despite favorable outcomes for many children born at the limit of viability, these children are at risk for mental health problems, with poorer school results.
Key Words: extremely immature • mental health • social competencies
Abbreviations: EI—extremely immature • ELBW—extremely low birth weight • VLBW—very low birth weight • ADHD—attention deficit/hyperactivity disorder • CBCL—Child Behavior Checklist • TRF—Teacher Report Form • DSM-IV-R—Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Revised • DSRS—depression self-rating scale • SES—socioeconomic status • MANCOVA—multivariate analysis of covariance • ANCOVA—univariate analysis of covariance • ES—effect size • OR—odds ratio • NSI—neurosensory impairment • CP—cerebral palsy
In the 1990s, major advances in perinatal intensive care and the development of regionalized perinatal care resulted in dramatic increases in the survival rates for extremely immature (EI) children born before 26 completed weeks of gestation.1–3 Follow-up studies of school-age outcomes for infants born in the 1980s with extremely low birth weight (ELBW) (<1000 g) showed that these infants had significantly high prevalence rates of low-severity neuropsychological, behavioral, and school difficulties.4–7 In a meta-analysis of case-control studies reported from 1980 to 2001 in which cognitive and behavioral outcomes were examined, it was found that school-aged children who were born very preterm exhibited more internalizing, attention, and externalizing problems, although these difficulties were not encountered in all studies.8 The vast research on behavioral outcomes of children with very low birth weight (VLBW) (<1500 g) or ELBW has revealed that these children are particularly vulnerable to problems related to inattention, hyperactivity, and social difficulties.9–19 There are also reports that preterm children with attention-deficit/hyperactivity disorder (ADHD) have significantly worse performance on cognitive tasks and a higher rate of school failure than control subjects.19–21 Furthermore, there is some evidence to suggest that, compared with control subjects, VLBW or ELBW adolescents12–14,19 and VLBW young adults22 have high prevalence rates of anxiety and depression. Apart from 1 report from Australia23 and an abstract from the United States,24 we are not aware of reports of behavioral outcomes at school age for children who were born extremely preterm (<28 weeks of gestation) or with ELBW in the 1990s. It is important to establish the mental heath of EI children born in the 1990s, because the dramatic increase in their survival rates in that decade might have increased the number of children at risk for adverse long-term outcomes.
In this part of our ongoing longitudinal investigation, the objectives were to examine behavioral outcomes, social competencies, and adaptive functioning in a national cohort of 11-year-old EI children, in comparison with matched, term, normal, control subjects, born in the 1990s and to investigate whether any observed differences could be explained in part by socioeconomic and environmental factors. We also addressed the question of whether correlates of the behavioral outcomes differed between EI children and control children. On the basis of the available literature, we hypothesized that the EI children would have high rates of ADHD and have social and possibly anxiety/depression problems, that there would be no significant difference in aggressive and delinquent behaviors between EI children and control children, and that the relationship between extreme immaturity and behavioral problems might be partially explained by psychosocial risk factors.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The control group was recruited at the present assessment (mean age: 11.6 years) from the national birth register, through selection of a child with a normal birth weight born at term at the same hospital, of the same gender and nearest in birth date (±7 days) to the EI child. We identified 3 control participants for every index child. Because we aimed to have 1 control child for every EI child, initially we contacted the first of the control families. If the family did not respond or refused to participate, then we approached the second family and, if necessary, the third. Recruiting the control families was a slow process, and this resulted in a control group that was an average of 8 months older than the EI children. Eighty-six percent (n = 74) of control families were recruited in the first round of invitation. Ten and 2 control families were recruited in the second and third rounds of invitations, respectively. The control group was approached and examined in the same way as the study population.
Instruments and Measures
Assessment of Mental Health
For assessment of the parents’ and teachers’ perceptions of the children's behavior, the parents completed the Child Behavior Checklist (CBCL) for ages 4 to 18 years28 and the teachers completed the analogous Teacher Report Form (TRF).29 Both forms include 118 items for scoring particular behavior/emotional problems, plus 2 open-ended problem items. The CBCL and TRF differ somewhat according to the home versus school context in which the child is seen, but the scoring profiles enable both instruments to be scored in terms of a common set of 8 problem behaviors. The list contains 118 items on difficult behaviors, all scored 0 (not true), 1 (somewhat or sometimes true), or 2 (very true or often true). Principal-component analyses reveal 8 sets of behaviors, namely, withdrawn, somatic complaints, anxious or depressed, social problems, thought problems, attention problems, delinquent behavior, and aggressive behavior. Principal-factor analyses of the 8 categories produce 2 broad groupings, namely, internalizing, derived from the sum of the items in the first 3 sets, and externalizing, derived from the last 2 (delinquent behavior and aggressive behavior). The remaining 3 categories (social, thought, and attention problems) represent problems that fit either broad grouping. The CBCL and TRF scales are also constructed in terms of items that measure 6 sets of behavioral syndromes based on the criteria of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Revised (DSM-IV-R),30 that is, affective disorder, anxiety, somatic problems, ADHD, conduct disorder, and oppositional-defiant disorder. Respondents were asked to base their answers on the preceding 6 months. To provide a common metric across the different tests, the raw scores for the CBCL and TRF problem subscales and the scores for the CBCL competence scales (activity, social, and school scales) were converted to z scores (SD scores). The z scores for the problem and competence scales of the CBCL were computed relative to Swedish normative values being used in Sweden at the time.31 Swedish reference normative values for calculation of z scores for TRF problem scales were not available; therefore, we computed z scores for TRF problem scales relative to the mean scores for the control subjects of the same gender. For all behavioral scales, higher scores indicated more problematic behavior. For all TRF and CBCL problem subscales, scores above the 90th percentile for the control subjects of the same gender were classified as being in the abnormal range. The percentile distribution of the total CBCL problem scores for our control group was similar to that for a Swedish reference population.31 Children completed a self-report with a depression self-rating scale (DSRS).32 The DSRS is an 18-item self-report questionnaire composed of a psychiatric symptom checklist that measures anxiety and depression. The child is asked to rate his or her own situation during the past month, on a 3-point scale. Scores of 2, 1, and 0 refer to most of the time, sometimes, and never, respectively. For the DSRS, scores above the 90th percentile for the control subjects of the same gender were classified as being in the abnormal range. The items of the DSRS were developed after extensive evaluation of the population and were tested against an operational clinical definition of depressive disorder.32
Measurement of Adaptive Functioning and Social Competencies
Parents provided responses for the following items from the competence scales of the CBCL: general activities (activity scale), which measure the amount and quality of the child's participation in sports, hobbies, and organized activities; social activities (social scale), which measure participation in organizations and clubs, number of friends and frequency of contacts with friends, and how well the child gets along with the parents, siblings, and other children; and school performance (school scale), which measures performance in academic subjects on a standard scale, grade retentions, any academic or other problems at school, and special education services received by the child at school. Assessments of academic performance and adaptive functioning at school were based on the TRF described by Achenbach.29 Teachers provided responses for the 5 items on adaptive functioning that measured the child's performance in academic subjects and 4 adaptive characteristics, as follows. How hard is she or he working? How appropriately is she or he behaving? How much is she or he learning daily? How happy is the pupil? School difficulties were defined as the child repeating a grade and/or using special educational resources (part-time or full-time).
Assessment of Demographic Characteristics
Sociodemographic characteristics, including variables such as the parents’ educational level, the family's disposable income, and the family structure, were assessed with the Nordic Health and Family Questionnaire.33 The mother's education was classified into 3 groups, that is, >12 years, 10 to 12 years, or 
9 years. The cutoff point for high/low education was between the latter group and the others. The family's disposable income was classified into 6 groups, and the cutoff point for high/low income was between the 2 highest groups and the others. The family structure was defined as a single-parent or 2-parent family. Any social risk was defined as single-parent family, mother's educational level of 9 years, or low family income. The composite social risk index included maternal education (0 for high school or above and 1 for 9 years of schooling), family structure (0 for 2-parent family and 1 for single-parent family), and family income (0 for high income and 1 for low income). The composite scale ranged from 0 for the lowest social risk index to 3 for the highest social risk. Socioeconomic status (SES) was measured on the basis of a Swedish socioeconomic classification system.34 In its most aggregated form, the classification consists of 6 groups. For the purpose of this study, we categorized SES into 3 main groups, that is, laborers (unskilled and semiskilled workers), nonmanual employees (assistant and intermediate nonmanual employees), and professional employees.
Chronic conditions at 11 years of age were defined as neurosensory impairments (NSIs) and medical or psychiatric illnesses with durations of 
12 months.25 Identification and characterization of NSIs in this study population were described previously.25 Chronic medical conditions were defined as NSIs and medical illnesses excluding psychiatric illnesses. Formal psychometric tests were not conducted but had been performed for all children who were receiving full-time special education (13 EI children and 4 control children). Major disability was defined as moderate or disabling cerebral palsy (CP), severe visual impairment (including unilateral or bilateral blindness), sensorineural disability requiring a hearing aid, or a need for full-time special education in a special school (as a measure of severe mental retardation).25
Assessment of Environmental/Family Risk
Family function was measured with a 12-item general functioning scale derived from the Nordic Health and Family Questionnaire.33 This instrument assesses family functioning in 5 dimensions, that is, affective responsiveness, problem-solving, affective involvement, communication and roles, and behavior control. The items concerning family functioning have 5 alternatives, graded from strongly agree to strongly disagree. Total scores range from 0 to 60, with higher scores indicating disturbance. Maternal mood and maternal psychological health were measured with a 12-item brief inventory devised by a research group in the Division of Psychiatry of Umeå University. On this scale, there are 6 items that measure maternal mood, that is, 3 items measuring positive affect (scores range from 0 to 18) and 3 items measuring negative affect (scores also range from 0 to 18). Low maternal mood was identified when the negative mood score was greater than the positive mood score. The positive mood affect was scored as 0 and the negative affect as 1. The remaining 6 items measure categorical responses to questions on the psychological health of the mother (visiting a psychiatrist or a psychologist, suicidal ideation, use of psychopharmacological drugs, phobias, and fears). Maternal health risk scores included all categorical responses from the maternal mood scale (0 for positive and 1 for negative maternal mood) and the psychological health scale (0 for not true and 1 for true). These scores were added to obtain a composite maternal mental health risk index (ranging from 0 to 12), with higher scores indicating disturbance. The study was approved by the regional ethical review board at Umeå University.
Statistical Analyses
Data were collected on standardized forms and encoded for computerized analysis with the use of Windows SPSS 13.0 (SPSS, Chicago, IL). The assessment data for each EI child were examined before they were combined with the data set from 2 previous main studies26,27 for analysis. Descriptive statistics such as frequency distributions, means, and SDs were used. Differences in dichotomous outcomes between the groups were analyzed with 
2 tests or Fisher's exact test, when appropriate. Continuous outcome measures were compared with unpaired Student's t tests, to test the differences between the means. Behavioral/emotional problems and competencies are likely to be related to stresses in a child's life, as well as factors such as social risk; therefore, in the analyses of the CBCL and TRF competence and problem scale scores, we controlled for social risk factors, family function, maternal mental health risk, and presence of a chronic medical condition in the child. Multivariate analysis of covariance (MANCOVA) was performed to test the differences between the groups. The independent variables were group status (EI versus control) and gender, and the dependent variables were each of the 6 DSM-IV-R-oriented syndrome scales of the CBCL or analogous scales of the TRF. The following variables were included as covariates/factors in the model: composite social risk index, family function, maternal mental health risk score, and presence of a chronic medical condition. MANCOVAs of CBCL and TRF syndrome scales were performed in 2 separate analyses. Follow-up univariate analysis of covariance (ANCOVA) was conducted for each of the 6 DSM-IV-R-oriented behavioral syndrome subscales of CBCL and TRF, with the same covariates as in the MANCOVAs of CBCL and TRF syndrome scale scores. ANCOVAs were also performed on the 5 TRF adaptive functioning scales (academic performance, daily learning, hard work, classroom behavior, and being happy), the 3 competence scales of CBCL (social, activity, and school scales), and the DSRS, with the same covariates/factors as in the ANCOVAs of the CBCL and TRF syndrome scale scores. The effect size (ES) is given by the partial 
2 statistic, which describes the proportion of total variability attributable to a factor or covariate (the proportion that, if multiplied by 100, is the percentage of total variability attributable to the group differences). All significant MANCOVA and follow-up ANCOVA effects were interpreted by using the criteria for ES described by Cohen,35 with which effects are deemed small, medium, and large if they account for 1% to 5.8%, 5.9% to 13.8%, and >13.8% of variance, respectively.
Multivariate logistic regression analyses were also performed to examine differences in dichotomous behavioral outcomes between the groups. Social risk score, family function, gender, maternal mental health risk score, and presence of a chronic medical condition were entered as covariates. P values of <.05 were considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Demographic Features
Sociodemographic characteristics, infant data, and rates of chronic conditions are shown in Table 1. At 11 years of age, 13 EI children (15%) had NSIs, which included 1 of the following conditions: CP for 5, severe visual impairment (including unilateral or bilateral blindness) for 10, and sensorineural disability requiring a hearing aid for 5. In the control group, the corresponding rate was 2% (n = 2; 1 child had CP, and 1 had severe visual impairment).25 Of the 86 EI children, 73 (85%) were in mainstream schools and 13 (15%) were receiving full-time special education. The corresponding rates for the control group were 82 (95%) and 4 (5%). The overall prevalence of 1 major disability was 21% for the EI children and 6% for the control participants (2 = 7.03; P = .006).25 There were no statistically significant differences between the EI and control participants regarding family structure, maternal education, maternal mental health risk index, SES, and family function (Table 1).
|
= .94, = .93, and = .84, respectively). The correlations among the behavior problem subscale scores from the CBCL and TRF ranged from r = 0.20 to r = 0.78 and from r = 0.18 to r = 0.75, respectively. Most of the subscales of the CBCL or TRF were correlated moderately with one another, with r values between 0.3 and 0.55. The mean ± SD Q correlation obtained between the CBCL item scores and the scores obtained for the corresponding item of TRF was 0.26 ± 0.19, indicating an average level of cross-agreement between teachers and parents.36
Behavioral Scores According to Parents’ and Teachers’ Reports
Parents and teachers reported significantly higher scores for EI children, compared with control subjects, for internalizing problems (anxious/depressed, withdrawn, or somatic problems) and attention, thought, and social problem scales. The mean z scores for EI versus control children according to the parent reports were 0.44 vs –0.17 (P = .002) for withdrawn behavior, 0.70 vs –0.14 (P < .001) for anxious/depressed behavior, 1.17 vs 0.15 (P < .001) for attention difficulties, 1.3 vs 0.39 (P = .001) for social problems, and 0.75 vs 0.1 (P = .024) for thought problems (Fig 1A). The pattern of findings was very similar for the teachers (Fig 1B). On the basis of both parents’ and teachers’ ratings, the mean raw total problem scores were increased significantly for EI children, compared with control children (CBCL score: EI: 23.1; control: 13.72; mean difference: 9.36; P = .001; TRF score: EI: 25.89; control: 15.05; mean difference: 10.84; P = .001). Compared with those of control children, parents and teachers of EI children were more likely to rate the child as scoring in the abnormal range for a number of behaviors (defined as CBCL and TRF problem scale scores above the 90th percentile for control children) (Table 2). The parent-reported rate of abnormal behavior among EI children for the anxious/depressed problem subscale was 27% (adjusted odds ratio [OR]: 2.56; P = .036), that for withdrawn problems was 36% (OR: 2.9; P = .011), and that for attention problems was 30% (OR: 3.5; P = .007). Teachers, like the parents of EI children, reported significantly higher rates of abnormal scores for internalizing (anxious/depressed or withdrawn) and attention problems for EI children, compared with control children. Teachers also reported significantly higher rates of abnormal scores for EI children for somatic, thought, and social problems (Table 2).
|
|
|
= 0.88; F6,152 = 3.54; P = .003; ES = 12%; social risk: Wilks’ = 0.90; F6,152 = 2.77; P = .014; ES = 10%; family function: Wilks’ = 0.91; F6,156 = 2.31; P = .036; ES = 8.2%; presence of a chronic medical condition: Wilks’ = 0.92; F6,156 = 2.25; P = .042; ES = 7.9%). No interactions emerged, which indicated that prematurity, family function, social risk, and presence of a chronic medical condition were associated independently with the multivariate measure of CBCL syndrome scales.
MANCOVA Effects According to Teachers’ Reports (TRF)
As with the parent reports (CBCL), MANCOVA of DSM-IV-R-oriented analogous TRF syndrome scales revealed a significant multivariate effect for prematurity (Wilks’ = 0.88; F6,152 = 3.47; P = .003; ES = 13%), which indicated that, across the 6 syndrome scales, there was a significant difference between the EI and control children. There was also a significant multivariate main effect for gender (Wilks’ = 0.85; F6,152 = 4.46; P < .001; ES = 15%). No interactions emerged, which indicated that prematurity and gender were associated independently with the multivariate measure of TRF syndrome scales.
Univariate Effects According to Parents’ Reports
Group Status (Prematurity Versus Control)
In the follow-up univariate analyses, all of the categorical independent variables and covariates/factors were the same as used in the MANCOVAs of DSM-IV-R-oriented syndrome scales of the CBCL and TRF. The analyses revealed significant effects of prematurity on 3 (affective problems, anxiety, and attention difficulties) of the 6 DSM-IV-R-oriented syndrome scales of the CBCL (Table 3), which suggested that the significant behavior problems persisted in these domains among the EI children after controlling for psychosocial risk factors, gender, and the presence of a chronic medical condition. The analyses also demonstrated that, in addition to prematurity, a number of other variables were associated significantly with behavior problems, as described below.
Family Function
Effects of family function emerged for the following domains: anxiety problems (F1,161 = 8.87; P = .003; ES = 5.2%), somatic problems (F1,161 = 4.61; P = .03; ES = 2.8%), and attention problems (F1,161 = 4.03; P = .046; ES = 2.4%).
Gender
Gender effects emerged for ADHD problems (F1,161 = 6.2; P = .014; ES = 3.7%), indicating that boys had higher ADHD scores than girls. However, there was no interaction between status and gender (F1,161 = 0.541; P = not significant), which suggested that EI boys and girls differed from their respective control subjects to similar extents.
Chronic Medical Condition
The presence of a chronic medical condition was associated with affective problems (F1,161 = 9.82; P = .002; ES = 5.8%) and anxiety problems (F1,161 = 6.32; P = .013; ES = 3.8%).
Univariate Effects According to Teachers’ Reports
ANCOVAs of 6 DSM-IV-R-oriented syndrome scales of the TRF revealed significant effects of prematurity on 4 disorders, namely, affective, anxiety, somatic, and attention difficulties (Table 3). As with the CBCL, a significant effect of gender on ADHD problems emerged (F1,158 = 20.23; P < .001; ES = 11.4%) and there was no interaction of group status with gender (F1,158 = 1.835; P = not significant). No other covariates were associated with any of the DSM-IV-R-oriented syndrome scales of the TRF. In a series of logistic regression analyses based on parent- and teacher-reported abnormal behavioral outcomes (CBCL and TRF), we found that adjustment for demographic variables (social risk, family function, maternal mental heath, presence of a chronic medical condition, and gender) made no material difference to the results, which remained significantly high in the EI cohort (Table 2).
Adaptive Functioning and Social Competencies
Comparisons between the EI and control cohorts were made with respect to a number of items that reflect adaptive functioning and social competencies. On the TRF scale that directly assesses academic performance in a mainstream school, EI children had significantly poorer scores than control participants (P < .001) (Table 4). ANCOVA revealed that the group status (prematurity) represented a large effect, accounting for 22% of the variance, which indicated that prematurity was associated strongly with poor academic performance, as indicated by teacher ratings. None of the other covariates was associated with the academic performance of either EI or control children. Furthermore, the EI children had significantly poorer scores than the control children with respect to TRF ratings of how much the child was learning (P < .001) and total adaptive function, which was computed by summing ratings for 4 adaptive characteristics (P < .001) (Table 4). There were no statistically significant interactions between gender and group regarding these variables. On the CBCL school and social scale, the EI children obtained poorer scores than the control participants (Table 4). Social risk was associated with the CBCL activity scale (F1,162 = 11.29; P = .001; ES = 6.5%). Although social risk was also associated with the CBCL social scale, it represented only a small effect (4% of the variance), compared with the medium effect (8%) observed for group status (prematurity versus control). Compared with control children, significantly larger proportions of EI children had failed a grade (15% vs 5%; 2 = 6.27; P < .05) or had school difficulties (59% vs 12%; 2 = 41.23; P <. 001) (Table 5). Thirteen EI children, compared with 4 children in the control group (P = .023), were receiving full-time special education in a special school or special class. A significantly smaller proportion of EI children, compared with control children, participated in any sports at school or at home (76% vs 92%; 2 = 7.38; P = .004) and significantly more EI children were rated below average in sports (29% vs 5%; 2 = 17.53; P < .001). Of the 21 EI children who did not participate in sports, 11 children had NSIs (CP for 4, visual difficulties for 6, and deafness for 1), 7 tired easily because of poor motor skills and poor coordination, and 3 had ADHD problems.25 However, no differences were found in the proportions of children who participated in activities and games other than sports. Furthermore, the proportions of children with impaired family relationships or relationships with peers and those who had few friends did not differ between the 2 groups (Table 5).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Studies of prematurity-related behavioral outcomes during adolescence for children born before the 1990s disclosed a wide array of emotional problems and behavioral disturbances.12–16,19,40,41 ADHD11–15,18,19,23,40–44 and internalizing,14,18,23,42 externalizing,15,18,42,43 social,40,45 and thought problems7,22 been linked to prematurity. In a meta-analysis, 227 studies reporting behavioral and cognitive data that were published between 1980 and 2001 were reviewed.8 Among 16 of those studies, presenting behavioral data that were considered worthy of review and involving children born before 1990s, 69% revealed significant increases in internalizing problems, 75% significant increases in externalizing problems, and 67% higher rates of attention problems. In the only population-based study on neurobehavioral outcomes of 8-year-old ELBW children who were born in the 1990s, the children were found to have attention difficulties, internalizing problems, and fewer adaptive skills, compared with the normal birth weight cohort.23 In our study, both the parents and teachers reported increased internalizing and attention difficulties and significantly lower scores for adaptive functioning for EI children, compared with control children.
In an international study of 408 ELBW children, 8 to 10 years of age, from 4 countries, behavioral problems were compared with a cross-cultural perspective7; behavioral data for all survivors from 4 large prospective studies of preterm children were obtained with the same instrument, namely, the CBCL. It was found that, despite cross-cultural differences, some behavioral problems and characteristics were strikingly similar for all cohorts. The mean scores for social, thought, and attention difficulty scales were 0.5 to 1.2 SDs higher than country-specific normative values or control values. In none of the cohorts were there any significant differences in the aggressive or delinquent behaviors between the ELBW children and the control children or a country-specific reference group. These findings were strikingly similar to our results; in our EI cohort, the mean z scores for social, thought, and attention problem scales, as rated by parents, were 0.75 to 1.3 SDs higher than those for the control group (Fig 1C). Similar ratings were reported by teachers (TRF) for our EI children. Furthermore, our EI children did not display externalizing (aggressive or delinquent) behavior problems, as indicated by parent or teacher report. This finding of an increase in attention/hyperactivity problems but not in conduct disorders or aggressive behaviors is in line with the findings from a number of other studies on behavioral outcomes for ELBW or VLBW children investigated at 8 to 12 years of age,7,9,12,14 in adolescence,19,41 or at a young adult age.22,46 It is particularly surprising that these children, despite having many risk factors for externalizing and risk-taking behavior (ADHD problems, social problems, and learning difficulties), do not have high rates of conduct disorders, according to teachers or parents. Furthermore, these results are not in agreement with those of follow-up studies of children with ADHD problems in the general pediatric population, which showed high rates of continuing psychopathological conditions and comorbidity of conduct and oppositional-defiant disorders.47 It has been postulated that the low rates of risk-taking may be attributable to increased parental monitoring and protection.46,48 An increased risk of anxious/depressed and withdrawn behaviors, as seen in our EI children, might lead to behavioral inhibition and to decreased antisocial behaviors.49 There is some evidence that the ADHD symptoms reported among VLBW children are more of the inattention type than of the hyperactivity type12 and are less associated with comorbidity and conduct disorders.9,12,14 This might also explain why ADHD symptoms in preterm children do not seem to have negative implications for the persistence of major sequelae into adulthood.50 However, it is evident that the ADHD profile seen in our EI cohort and in many other VLBW or ELBW populations does not represent a clinical diagnosis of ADHD. Others have suggested that the interaction between perinatal and social risk factors may be a critical factor.51 More studies are needed to elucidate how child outcomes are influenced by the interaction between parenting and perinatal complications. However, the mental health outcomes in adulthood for extremely preterm children born in the modern era of perinatal care remain to be seen.
The depression rating scale includes items such as "life not worth living,""feels sad," "lonely," "does not sleep well," "feels bored," and "cries a lot."32 Although they do not measure exactly the same construct, the anxious/depressed syndromes of the CBCL and TRF scales comprise similar questions.28,29 Our findings showed increases in anxious/depressed and withdrawn symptoms for EI children according to parents’ and teachers’ reports and a significant increase in the trend toward depression symptoms according to children's self-reports. Increased rates of depression,13,14,19,22,45 anxiety,13,42 and overall internalizing behaviors, including shy and withdrawn behaviors,44,45 have been reported for VLBW or ELBW children. A direct link between prematurity and isolated/withdrawn behaviors at school has been reported.45,52 According to the mediational model postulated by Nadeau et al,44 it seems that functional limitations that are consequent to the extreme prematurity may explain anxious/withdrawn and isolated behaviors. Middle school age is considered to be full of competitive tasks in a teenager's life; in particular, physical activities such as team and individual sports are highly valued. In our cohort, more than one fourth of the EI children had poor motor skills, one fourth were restricted in their activities because of 1 handicap, and almost one half had moderate/severe learning difficulties. These rates remained significantly high when children without NSIs were analyzed.25 We, like others,44 think that these difficulties might have imposed on our EI children functional limitations in social situations, leading to more-cautious (even anxious and withdrawn) behaviors during interactions with peers. However, in-depth studies are needed to determine the extent to which these limitations create handicaps for these children that lead them to withdraw from their social groups. Our findings are in conformity with those of others who reported that children with chronic medical illnesses were at increased risk for emotional problems, particularly depression/anxiety and peer conflict/social withdrawal.53–55 Furthermore, in our study, as found by others,13,16,19 differences in behavioral problems persisted whether children with NSIs were excluded.
The clinical interpretation of the increased scores for thought problems according to parent and teacher reports in our EI cohort is not well understood. Similar results were obtained in a multinational study of 8- to 10-year-old ELBW children7 and by Hack et al,22 who found higher rates of thought problems for both VLBW men and women, compared with control subjects, in a behavioral study of young adults at the age of 20 years. The CBCL and TRF thought problem scales are relatively short subscales and reflect a heterogeneous group of DSM-IV-R disorders.56 They include items such as "can't get mind off," "repeats acts," "sees things," "hears things," "strange behavior," and "strange ideas," with the TRF having additional items such as "harms self" and "fears." An association between schizophrenia and low birth weight has been suggested from a few epidemiological studies, but it is not clear whether this pertains to intrauterine growth failure, prematurity, or both.57–59 There is some evidence suggesting an association between autistic spectrum disorder and prematurity.60 Others have reported an increased incidence of autism in very preterm children with an unfavorable visual status attributable to severe retinopathy of prematurity.61,62 The implications of these behavioral difficulties reported by both parents and teachers for our EI children are not clear. It is likely that these behavioral tendencies will decrease with age as the children grow, as do other problems such as hyperactivity15 and externalizing behavior.63
Sociodemographic and environmental factors also significantly influence the long-term outcomes and quality of life for children born preterm.13,19 Saigal et al19 found that birth weight, family function, maternal mood, gender, and SES predicted the behavioral adjustment of immature adolescents. Levy-Shiff et al13 reported that birth weight and psychosocial/environmental variables were significant predicators of later emotional adjustment. McCormick et al17 assessed the contributions of birth weight status and family environment to behavior problems at school. They pointed out the greater importance of environmental factors, compared with birth status, and proposed interventions aimed at reducing the environmental risks. Klebanov et al64 found that, beside birth weight status, maternal mental depression and home environment were associated significantly with social competence scores. In our study, boys showed a significantly increased risk for attention problems. Family function, social risk, and presence of a chronic medical condition predicted behavioral adjustment in our study population. Although we did not measure exactly the same constructs, our findings were similar to those of Levy-Shiff et al,13 Saigal et al,19 and Klebanov et al64 with respect to the importance of family environment and social risk in predicting the behavioral adjustment of preterm children. Furthermore, in accordance with previous studies,12,19 we did not find any interactions between group status (EI and control groups) and other covariates, which indicated that the correlates of behavioral outcomes did not differ between our EI children and the control children.
In terms of school performance, as reported by parents and teachers, more than one half of the children in our EI cohort were experiencing school problems. However, 85% of the EI children were attending mainstream schools and the majority of them were not having major adjustment difficulties. Despite there being fewer adaptive skills in our EI cohort, these children were not different from the control children with respect to being happy and being positively adjusted in their daily lives. Although the presence of a chronic medical condition was associated with poor school performance, as reported by parents, it represented only a small effect (3%), compared with the large effect (16%) of group status (ie, of being born extremely preterm). No other environmental variables were associated with school problems, as evident from parents and teachers. ADHD problems were a significant predictor of school difficulties, as reported by parents and teachers (data not shown). Results pertaining to executive functions/ADHD problems and school achievement will be the subject of a separate report. Our findings are in agreement with previous reports that showed higher rates of school failure and of grade retentions for ELBW teens, compared with control subjects.16,19,65–68 Furthermore, our results are similar to those of a number of other reports that demonstrated significant associations between ADHD and poor academic performance, high rates of school failures (including lower grades at school), and receiving full-time special education.11,19–21,66–68 In terms of adaptive functioning, as judged from parent reports, our EI children were seen by their parents as being less well-adjusted in extracurricular activities. These findings could not be explained in terms of more-disabled children who were unable to participate, because these differences persisted when children with NSIs were excluded from the analysis. Others have reported a similar finding of poorer social adjustment, involving mainly social functioning as judged by social attainments and global evaluations of competence, especially in activities such as sports.19,41 However, more than two thirds (76%) of our EI children were participating in coached sports, and no difference was seen between the groups in having few friends or in teacher and peer relationships.
Various plausible mechanisms might explain the association between extreme immaturity and behavioral problems. MRI studies of the brain in immature children have shown a relationship between an abnormal brain structure and certain behavioral difficulties.69,70 Perinatal and neonatal complications that lead to central nervous system insults have been shown to be related to ADHD problems.10,69 Nagy et al71 reported brain MRI findings for 27 preterm children with ADHD problems at 11 years of age; the children had not suffered from intraventricular hemorrhage or periventricular leukomalacia in the neonatal period. It was found that the preterm cohort had significant white matter disturbances and reduced brain volumes, compared with control subjects. There is good evidence that disruptions in cortical development (corticogenesis) and brain connectivity are related inversely to birth weight and gestational age at birth, particularly for very preterm infants, even in the absence of a concomitant biomedical risk.72,73 Despite increased knowledge regarding these processes, the functional significance of brain abnormalities is not clear. Social isolation, that is, difficulty in maintaining social relationships and lack of ability to make friends, has been reported to occur more frequently in children who are hyperactive or have attention-deficit disorder.74 Life-threatening events in the neonatal period might result in increased parental monitoring and overprotective parental attitudes, leading to deviations in perceptions of the child's behavior and to inadequate socioemotional adjustment in the child.75
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
We are indebted to the study children and their families and schoolteachers for their cooperation. We especially thank Dr Hans Stenlund in Umeå for his statistical advice and expertise. We also thank research nurse Margareta Backmän (Umeå) and project assistant Nighat Farooqi (Umeå) for collecting data and establishing invaluable contacts with the families.
|
FOOTNOTES |
|---|
Address correspondence to Aijaz Farooqi, MD, PhD, Division of Pediatrics, Department of Clinical Sciences, University Hospital, SE-901 85 Umeå, Sweden. E-mail: aijaz.farooqi{at}pediatri.umu.se
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
800 g) adolescents who are free of major impairment, compared with term-born control subjects.
Pediatrics. 2004;114(6)
. Available at: www.pediatrics.org/cgi/content/full/114/6/e725
